Archive

Archive for January, 2014

Weekly QUEST Discussion Topics and News, 31 Jan

January 31, 2014 Leave a comment

TITLE: QuEST for DDDAS through Information Fusion —- Erik Blasch

ABSTRACT QuEST (QUalia Exploitation of Sensing Technology) has recently been focused on a theory of knowledge for situation understanding as supporting the decision quality between a computer and human agent. To achieve QuEST, numerous cross-discipline approaches are needed of which two include information fusion and dynamic data driven application systems (DDDAS). DDDAS brings together theoretical modeling/simulation, instrumentation measurements, applications algorithms and systems software. As noted then, traditional approaches to information and QuEST have developed many methods for measurement and algorithms; however, the theoretical and systems software are needed for future systems. In this talk, we (1) focus on the relations between DDDAS, Information Fusion, and QUEST, (2) demonstrate a use case of importance to the QuEST team in a user-defined operating picture, and (3) pose challenges that surround the combined approaches. The talk is supposed to be in the spirit of the QuEST discussions in bringing a few ideas and posing a research collaboration within the team.

References
DDDDAS Program Review

E. P. Blasch, E. Bosse, and D. A. Lambert, High-Level Information Fusion Management and Systems Design, Artech House, Norwood, MA, 2012.
E. Blasch, “Enhanced Air Operations Using JView for an Air-Ground Fused Situation Awareness UDOP,” AIAA/IEEE Digital Avionics Systems Conference, Syracuse, NY, Oct. 2013.
E. Blasch, J. J. Salerno, I. Kadar, S. J. Yang, L. H. Fenstermacher, M. Endsley, L. L. Grewe, “Summary of human social, cultural, behavioral (HSCB) modeling for information fusion panel discussion,” Proc. SPIE, Vol. 8744, 2013.
E. Blasch, “Book Review: 3C Vision: Cues, Context, and Channels,” IEEE Aerospace and Electronic Systems Magazine, Vol. 28, No. 2, Feb. 2013.
….. And associated references within the above

Other topics brought up this week of interest include and potential topics for future meetings:
Autobiographical memory article in Feb issue of Sci Am
Stockham – 1972 IRE article
Computing with words – NLU is NOT equal to computing with qualia slides 63 …
Thinking chapter provided by Robert P
More on the NLU article from last week

WeeklyQUESTDiscussionTopicsandNews31Jan

WeeklyQUESTDiscussionTopicsandNews31Jan

Weekly QUEST Discussion Topics and News, 24 Jan

January 23, 2014 Leave a comment

As always we welcome any discussions people desire to engage in – for example there have been several email trails associated with our what is QuEST document:
Jargon based answer:
QuEST (what was the acronym?) QUalia Exploitation of Sensing Technology – QuEST – a cognitive exoskeleton

• QuEST defines and engineers a new set of processes that will be implemented in a computer agent (or set of computer agents) to improve decision quality of a human agent (or set of human agents)

• Assumption 1: decision quality is dominated by the appropriate level of situational awareness (The perception of environmental elements with respect to time and/or space and/or logical connection, the comprehension of their meaning, and the projection of their status after some variable has changed, such as time.)

• QuEST could be considered a new approach to situational assessment (processes that are used to achieve situational awareness), situation understanding (comprehension of the meaning of the information as integrated with each other and in terms of the individual’s goals. It is the “so what” of the data), or sensemaking (a motivated, continuous effort to understand connections which can be among people, places, and events in order to anticipate their trajectories and act effectively” Klein et al., 2006, p. 71) for decision quality

– QuEST agents implement blended dual process cognitive models (have both artificial conscious and artificial subconscious/intuition processes) for situational assessment

– The artificially conscious processes all are constrained by the fundamental laws of the QUEST Theory of Consciousness (structural coherence, situation based, simulation / cognitively decoupled)
– The subconscious/intuition processes do not use working memory and thus considered autonomous (do not require consciousness to act) –current approaches to Data driven artificial intelligence provide a wide range of options for implementing instantiations of capturing experiential knowledge used by these processes.

• QuEST is developing a ‘Theory of Knowledge’ – to provide the theoretical foundations to understand what an agent or group of agents can know which fundamentally changes machine learning from an empirical effort to a scientific effort

Street speak answer:

• QuEST improves decision quality by providing decision makers computer based decision aids that are engineered with both intuition and the ability to do deliberative thinking

• QuEST seeks a mathematical framework to understand what can be known by a person or group of people and their computer based decision aids about situations so we can predict when more people (or differently trained people) or more information are necessary to make a particular decision.

WeeklyQUESTDiscussionTopicsandNews24Jan

WeeklyQUESTDiscussionTopicsandNews24Jan

Weekly QUEST Discussion Topics and News, Jan 17

January 17, 2014 Leave a comment

The topics for this week include a discussion on the ideas to answer the question ‘What is QuEST?’ – below is some draft text based on some feedback.
Jargon based answer
QUalia Exploitation of Sensing Technology – QuEST – a cognitive exoskeleton

• QUEST defines and engineers a new set of processes that will be implemented in a computer agent (or set of computer agents) to improve decision quality of a human agent (or set of human agents)

• Assumption 1: decision quality is dominated by the appropriate level of situational awareness

• QUEST could be considered a new approach to situational assessment (processes that are used to achieve situational awareness), situation understanding, or sensemaking (depending on the application) for decision quality over short term versus long term, tactical vs. strategic, and individual versus group conditions

– QUEST agents implement blended dual process cognitive models (have both artificial conscious and artificial subconscious/intuition processes) for situational assessment (processes that are used to achieve situational awareness)

– The artificially conscious processes all are constrained by the fundamental laws of the QUEST Theory of Consciousness (structural coherence, situation based, simulation / cognitively decoupled)
– The subconscious/intuition processes are processes that do not use working memory and thus considered autonomous (do not require consciousness to act) – we believe current approaches to Data driven artificial intelligence provide a wide range of options for implementing instantiations of capturing experiential knowledge.

• QUEST is developing a ‘Theory of Knowledge’ – to provide the theoretical foundations to understand what an agent or group of agents can know which fundamentally changes machine learning from an empirical effort to a scientific effort

Acknowledgement: Over decades of research in perception and recognition from single sources; the information age opens up a distributed set of data, users, and decisions which requires a research group from varied backgrounds to answer fundamental questions of a “situation” for different “agents”.

Street speak answer:

• QUEST improves decision quality by providing decision makers computer based decision aids that are engineered with both intuition and the ability to do deliberative thinking to match results with needs

• QUEST seeks mathematical foundations to understand what can be known by a person or group of people and their computer based decision aids about situations so we can predict when more people (or differently trained people) or more information are necessary to make a particular decision.

The second area for discussion to keep with our goal of weekly exposing the group to an article or area of literature that may be key to delivering QuEST agents – we have two articles we want to vector people to:

Cogn Process (2010) 11:103–121
Good judgments do not require complex cognition
Julian N. Marewski • Wolfgang Gaissmaier •
Gerd Gigerenzer
What cognitive capabilities allow Homo sapiens
to successfully bet on the stock market, to catch balls in
baseball games, to accurately predict the outcomes of
political elections, or to correctly decide whether a patient
needs to be allocated to the coronary care unit? It is a
widespread belief in psychology and beyond that complex
judgment tasks require complex solutions. Countering this
common intuition, in this article, we argue that in an
uncertain world actually the opposite is true: Humans do
not need complex cognitive strategies to make good
inferences, estimations, and other judgments; rather, it is
the very simplicity and robustness of our cognitive repertoire
that makes Homo sapiens a capable decision maker.
Cogn Tech Work (2005) 7:14-28
Gary Klein Æ Rebecca Pliske Æ Beth Crandall
David D. Woods
Problem detection
Abstract Problem detection is the process by which
people first become concerned that events may be taking
an unexpected and undesirable direction that potentially
requires action. Previous accounts [e.g., Cowan (Acad
Manage Rev 11(4):763–776, 1986)] described problem
detection as the accumulation of discrepancies until a
threshold was reached. In reviewing incidents taken
from a variety of natural settings, we found that discrepancy
accumulation did not apply to the incidents we
reviewed, because (a) cues to problems may be subtle
and context-dependent, and (b) what counts as a discrepancy
depends on the problem-solver’s experience
and the stance taken in interpreting the situation. In
many cases, detecting a problem is equivalent to reconceptualizing
the situation.

WeeklyQUESTDiscussionTopicsandNewsJan17

WeeklyQUESTDiscussionTopicsandNewsJan17

Weekly QUEST Discussion Topics and News, 10 Jan

January 9, 2014 Leave a comment

We want to continue our discussion on the current view of the QUEST overview – ensure we answer any/all questions on what to date we’ve concluded – including an update on our recent interactions with DARPA / AFOSR / ONR and the AF/ST chief scientist on the overview talk. The discussion will include details on what needs to be done next to advance our understanding of the issues and how we might engineer solutions as well as advance the foundational framework. One specific topic is how to instantiate in a LaRue-like model a representation / deliberative processes that are consistent with our ‘Theory of Consciousness’.

The second topic is a follow on to the above discussion – we want to ensure every week we expose the group to new sources of ideas that will help us mature our positions and also help us understand where we fit in the trade space. This week I want to expose the group to the chapter provided to us by Mike Young late last year on Dual Process Theories – Betram Gawronski and Laura Crieghton (2013), D. E Carlston (Ed.) The Oxford handbook of social cognition )pp 282-312, Ny NY Oxford university press. The abstract is below – but it is a fascinating work that is somewhat a distinct set of literature as it focusses on social cognition models.

Abstract
Dual process theories divide the realm of mental processes into two general categories depending on whether they operate automatically or in a controlled fashion. This chapter provides an overview of dual process theories in social psychology, focusing on their historical and conceptual developments. Identifying three general categories of dual process theories, the chapter distinguishes between domain-specific theories that focus on particular phenomena, generalized dual process theories that identify domain-independent principles underlying various kinds of phenomena, and formalized dual process theories that quantify the joint contributions of automatic and controlled processes to responses within a single task. The chapter also discusses critical arguments against each type of dual process theorizing, which are integrated in a general outlook on future directions.
Key Words: attitudes, attribution, automaticity, control, dual process theories, impression
formation, persuasion, prejudice, stereotyping

WeeklyQUESTDiscussionTopicsandNews10Jan

WeeklyQUESTDiscussionTopicsandNews10Jan

Weekly QUEST Discussion Topics and News, 3 Jan

January 2, 2014 Leave a comment

Kabrisky Memorial lecture on the ‘State of Quest’ 2014

Situationally conscious cognitive Exoskeleton – QUEST
1.) Will explain later the idea of situational consciousness vs situational awareness – but the concept of a cognitive exoskeleton is the key thrust of this effort – how to deliver decision advantage – improving decision quality is what we seek
2.) Most suggest that (decision advantage = improving decision quality) is achieved through SA – data like that provided by AF/ST support and the aviation slide – SA drives decision quality (second AF/ST slide) – define SA (wiki)– must be general enough to be useable in Cyber next cyber slide – ignores people – but the layered view from AF/ST includes – Lots of work (for example by AF/ST) that conclude that decision quality is dominated by SA (situational awareness)- last cyber slide captures some of the key cyber challenges – So is the key function for a cognitive exoskeleton to ensure SA?
3.) Although this effort / discussion is focused on ways to impact C4ISR in and across all domains – the approach is very general and we’ve investigated for example applications in sustainment of the fleet, cyber ISR, etc. – the AF/ST slide goes through the need for autonomy across our core functions – virtually anywhere ‘decision’ quality is the limiting factor in capability – and that is everywhere. – jeff Jonas view of problem space – Endsley equivalent to Jonas view – lots of data but little information needed
4.) Sensor designers concluded that need more and more sensors to achieve SA (AF/ST slide on information gap and why? – and the result is lots of stove piped solutions each bringing a piece of the data space that could improve SA to a human that must integrate and filter –
5.) Cognitive engineering (user centered philosophy) and SA oriented design slides – provides insight into a new approach to solve these issues – side note – we’ve successfully used these cognitive engineering approaches in our ongoing pcpadx efforts to revolutionize the DCGS for SOF etc – unifying frame of reference was a situation! Those slides are in the back up area – fyi for AF/ST
6.) Cognitive Engineering insight also provides a new view of what SA is – define Situations – Asituation is any part the agent centric internal representation of an agent which can be comprehended as a whole by that agent through defining how it interacts with oris related to other parts of the representation in that agent. We will definecomprehended by defining how it interacts or is related to other situations via linking (and types of links). interacting with other things we mean that the situations have properties or relate to other situations.” *** we would say can and must be linked to other ‘situations’ = ‘other qualila’ = other chunks*** – agent centric – there are not out there – they are in the mind of the agent making the decision – thus the idea of measuring the world is NOT the solution to SA – sensing is –
7.) Sensing as a new paradigm – SoS is the key – sensors integrated within decision systems – idea is sensors as part of SoS that make decisions that is what makes sensing versus sensors – Tenets of Sensing as a Service – Embedding in closed-loop decision systems – Responsive sensing automation – Streaming analytics and multi-modal fusion – Accessible across domains – Multi-user support – Comprehensive assessment – the goal is to address the challenge presented in the Strategic vision ‘Delivering Decision Advantage’ – “AF ISR’s processing, exploitation, and dissemination (PED) capability has evolved considerably over the last decade. To continue the maturation, we will break the linear relationship between collection and analysis, where every increment of additional collection capacity requires a proportionate increase in analytical manpower. We will embrace the need for increased automation while recognizing that analysts play the critical role in synthesis, integration, and insight.” ** we would include also breaking the relationship between the number of ‘customers’ accessing the information and having their decisions impacted and the resources used **
8.) But in the modern battle spaces decision quality has to be improved in a range of decision makers – thus we need to break the linear relationship between … – thus we propose sensing as a service –customer interacting with the data with ToM –
9.) *** up to this point it is background – to provide context for AF/ST – can start with slide 19 for just a quest overview *** Most suggest solution lies in autonomy
10.) Magic bullet will not occur
11.) Need human machine teams – IA vs AI – AF/ST slide – her conclusion is justification – the question is how to achieve human/machine teaming – how to overcome the problems she list in her challenges to effective use of autonomy
12.) Joint cognitive systems – we suggest the solution lies in realizing human machine teaming within a common framework – with shared perception –
13.) Requires cognitive models – we’ve focused on dual process theory like advocated by Evans and Stanovich –
14.) Dual process ideas – start with type one processes and examples
15.) Rating professors and processing sensory data (example blindsight)
16.) Then the type two processes that lead to consciousness
17.) Define qualia – and define consciousness as generating qualia
18.) Key idea is artificial consciousness – that requires a definition / understanding of consciousness – what is it? – why is it there – qualia – cognitive flexibility – hypothetical part of SA levels is the key insight – but extend to not just be what might happen in the future but also what might be happening now and what might have happened before – imagined present, imagined past and imagined future – all consciousness –
19.) Requires a theory of consciousness – set of fundamental laws – situation based / simulation / structurally coherent
20.) First of laws is situational based – situated –
21.) Key missing piece of current AI is means to process situations – / perceptions
22.) Define situations
23.) Second law is cognitively decouple / hypothetical simulation = type 2 processes (consciousness)
24.) Related to Endsley model of SA
25.) We agree that the projection level of SA is a hypothetical – an imagined future – we think that idea is key – would like to extend to have a hypothetical present / hypothetical past – need computational theory of processing these situations / hypothetical situatiions –
26.) Define awareness – versus consciousness – range of values versus awareness
27.) Need approach to process situations / perceptions / chunks / qualia
28.) We seek quest agents to serve as cognitive exoskeletons – for human or groups of humans – their goal is to use a dual process representation to capture situations appropriate for their human team-mates –
29.) By having all (humans and computer based quest agents) using the same approach to representation we have hope of a common framework – a common math that can compute a situational complexity measure that can be used to determine what can be known by a group of agents – allowing calculation of capacity (situational complexity based)
30.) Requires a framework for the dual process models – example is the Stanovich and Evans ideas and a common mathematical framework (possibly category theory) capable of handling computer agents with artificial consciousness and humans that will lead to a theory of knowledge (what can be known by this set of agents using this set of sensing solutions)
31.) Requires models – example Larue (have early working versions now)
32.) Requires a set of driver problems – example pcpadx cp2 and cyber SA

WeeklyQUESTDiscussionTopicsandNews3Jan